Mastering Photogrammetry: Crafting 3D Models with RealityCapture | Unreal Fest 2024

preview_player
Показать описание
Join us for a practical introduction to photogrammetry using RealityCapture. In this session, we'll walk you through the process of creating 3D models from images, focusing on a straightforward workflow that can benefit professionals across various industries. We'll explore essential post-processing techniques like smoothing and simplification to enhance model quality and fit polycount budgets. You will also learn how to bake textures from one model to another and how to create normal and displacement maps. Finally, you will see how different types of models can be imported into Unreal Engine for a range of polycount and texture budgets. By the end of this session, you'll have a solid understanding of the photogrammetry workflow and be equipped to utilize RealityCapture effectively in your projects.
Рекомендации по теме
Комментарии
Автор

The depth mapping process is such a clever way to do it. Thanks for the information. @38:55

ScenegraphStudios
Автор

**Introduction and Overview:**
- **0:01**: Introduction by Wallace from the product team at Capturing Reality, focusing on photogrammetry and the Reality Capture and Reality Scan software.
- **0:19**: Overview of photogrammetry principles and common issues.

**Principles of Photogrammetry:**
- **1:00**: Scanning involves capturing many images from different angles to create a 3D model using software.
- **1:20**: Laser scans and camera rigs, as well as drones with RTK GPS, can improve accuracy.
- **2:08**: Key considerations: image quality, overlap, and subject coverage.
- **3:05**: Image quality and surface features are crucial; blurry images and noise can cause problems.
- **3:44**: Overlap is essential; 70% is a good rule of thumb, but more overlap can improve results.
- **4:34**: General photogrammetry process: capturing images, processing with Reality Capture, and exporting as a 3D model.

**Photogrammetry Equipment:**
- **5:01**: Equipment includes cameras (mobile phones, DSLRs), lenses, tripods, rulers, color checkers, lighting equipment, and polarization filters.
- **7:38**: Using drones for large objects, turntables for small objects, and scanning sprays for shiny objects.
- **9:05**: Additional gear includes smartphones, laptops for on-site processing, and practical items like batteries and power banks.

**Guidelines for Taking Photos:**
- **11:04**: Do not limit the image count; use high resolution; each point should be visible in at least two images; shoot full loops around the object; maintain high overlap; avoid panoramic images; avoid object movement; avoid transparent and shiny objects; maintain consistent lighting conditions.
- **12:47**: Importance of taking many photos and maintaining high resolution.
- **13:29**: Photography basics: aperture, shutter speed, and ISO settings are crucial for quality.
- **15:59**: Ensure each point is visible in at least two images; follow the shape of the object; maintain high overlap.

**Additional Tips:**
- **17:04**: Avoid changing the viewpoint more than 30°; keep the object in the frame; use high overlap to tie images together.
- **18:14**: Avoid tilting the camera; keep features visible in multiple images; avoid featureless textures and inconsistent lighting.
- **19:00**: Practical advice on dealing with featureless textures and maintaining consistent lighting.

**Dos and Don’ts:**
- **21:29**: Dos include using high-resolution cameras, consistent lighting, and proper tripod use. Don’ts include limiting image count, losing focus, capturing motion blur, and using transparent or featureless textures.


**Camera and Lens Recommendations:**
- **26:46**: Recommended lens focal lengths: 24mm (wide), 50mm (eye level), 100mm (telephoto). Ideal range is between 35mm and 50mm.
- **27:20**: Primes offer sharper images than zoom lenses. Dynamic range is important to capture details in both shadows and highlights.
- **27:49**: Use histograms to check exposure; underexposed images can lose detail in shadows, and overexposed images lose detail in highlights.
- **28:14**: Depth of field: Examples of images taken at F16 (long depth of field) and F2 (short depth of field). Both can be useful depending on context.

**Post-Processing Images:**
- **28:49**: Review and discard blurry or low-quality images before processing.
- **29:07**: Use tools like Lightroom for batch processing and adjusting exposure, shadows, and highlights.
- **29:59**: Color checkers help ensure accurate colors. Be cautious with sharpening filters to avoid unwanted artifacts.

**Alignment and Reconstruction:**
- **30:33**: Align cameras and merge different inputs (photos, laser scans) for the best result. Combining laser scans with photos can produce superior mesh and texture quality.
- **31:07**: Save XMS (camera positions) as draft or locked for alignment. The reconstruction stage uses GPU and can be time-consuming.
- **32:12**: Mesh editing: Use decimation to simplify large meshes, and bake normal maps from high-resolution models.

**Texturing and Exporting:**
- **33:29**: Calculate vertex colors and textures. Unwrapping tools help create textures with high detail. UDM (UDIM) format is used for detailed texturing.
- **34:29**: Export options include various model types, orthographic projections, and fly-through videos.

**Live Demonstration:**
- **35:09**: Live demonstration of scanning a jug, including setting up images, creating masks, and using Reality Capture for alignment and reconstruction.
- **37:25**: Post-processing includes creating masks and handling different orientations of the object.

**Practical Workflow:**
- **40:09**: Importing images and using masks to prevent confusion during scanning. Explanation of the scanning process with masks to ensure complete coverage.

**Export and Integration:**
- **43:22**: Demonstration of using the simplification tool to manage high-detail models and create low-poly versions with baked textures and normal maps.
- **45:02**: Exporting models to Unreal Engine using UDIM format. Setting up textures and materials in Unreal Engine.

**Closing:**
- **48:25**: Summary of the differences between importing detailed environments versus small objects. Final thoughts and wrapping up the presentation.

._
Автор

Can you post your power point presentation please.

cgrosbeck
Автор

50mm standard prime or 50mm macro? Thanks!

BubbaSatori
Автор

wow my mind just asked me if he was real even or 3d. because i just started to wonder what if they were testing people to see or show them you can't tell the difference between real life and mastered reality capture :D that light on half his face was part of what made me wonder. testing two different light patterns on the face to see if anyone could tell or guess. or maybe to mask or hide a slight problem with it? nah i'm probably just being... over imaginative. or am I

aurorauplinks
Автор

RC is a great app, one of the best.. it needs "only" a complete UI overhaul :)

dainjah